Diagnostically Lossless Compression Strategies For X Ray Angiography Images

Author : Emhemad Mohamed Saffor
Publisher :
ISBN 13 :
Total Pages : pages
Book Rating : 4.:/5 (975 download)

DOWNLOAD NOW!

Download or read book Wavelet-based Loosy Compression Techniques for Medical Images written by Emhemad Mohamed Saffor and published by . This book was released on 2003 with total page pages. Available in PDF, EPUB and Kindle. Book excerpt: Medical imaging is a powerful and useful tool radiologists and consultants, allowing them to improve and facilitate their diagnosis. Worldwide, X-ray images represent 60% of the total amount of radiological imagfes, the remaining consists of more newly developed images modalities such as Computed Tomography (CT), Magnetic Resonance Imaging (MRI), Ultrasiund (US), Positron Emission Tomography (PET), Single Photon Emission Computerized Tomography (SPECT), Nuclear Medicine (NM), and Digital Substraction Angiography (DSA). Image communication system foe medical images have bandwith and image size constrain that reseult in time-consuming transmission of uncompressed raw image data. thus image compression is a key factor to improve transmission spped and storage, but it risks losing relevant medical information. The radiology standard Digital Imaging and Communication in Medicine (DICOM3) provides rules for compression using lossless Joint Photographic Expert Group (JPEG) methods. However, at the moment there are no rules for acceptance of lossy compression in medical imaging and it is an ectremely subjective decision. Acceptable levels of compression should never compromise diagnostic information. Wavelet technology has emerged as a promising compression ratio while maintaining an acceptable fidelity of image quality. the objective of this thesis is to evaluate a variety of wavelet filters using Wavelet toolbox for selecting the best wavelet filter to be used in compressing and decompressing selected medical images. Two-dimensional wavelet decomposition, quantization and reconstruction using several families of filter banks were applied to a set of medical images. Furthermore, the technique of quantifying the effect of wavelet compression using low and high contrast test object on digitized radiographic chest and abdomen images and Computed Tomography (CT) images is discussed. A test trip simulating low and high contrast objects was constructed using constructed using Visual Basic programmning environment. Lurawave Smart Compression software was used to compress and reconstruct selected images with different compression ratios. Ten observes, from he Department of Radiology at the University on Malaya Medical Center (UMMC) evaluated the results. Objective and subjective methods were used for evaluating our results, which include Peak Signal to Noise Ratio (PSNR), Receiver Operating Characteristic (ROC) analysis and rank sum test. The result show that no specific wavelet filter perform uniformly better than others except for the case of Daubechies and bi-orthogonal filters, which are perhaps, the best wavelet filters for selected medical images. These filters, give very small (= zero) Maximum Absolute Error (MAE). These wavelet filters were able to produce perfect reconstruction, even if the level of wavelet decomposition increase. Using LuraWave Smart Compression software, the result show that PSNR which represent the quality of reconsructed images, was between 36 to 42 dB and 28 to 36 dB for chest and abdomen images respectively. All these images were compressed and decompressed up to 500 : 1. For CT images the PSNR was between 36 to 57 dB for chest image, 34 to 52 dB for abdomen images and 40 to 57 dB for CT brain images. All these images were compressed and decompressed up to 30 : 1 compression ratio. The ROC analysis, which represent the diagnostic quality of digitized chest X-ray images, indicated that compression ratio 100 : 1 is still acceptable for digitized chest and abdomen X-ray image, whereas for CT images, the result indicated that compression ratio 30 : 1 is acceptable for chest and abdomen images and 20 : 1 for brain image. The investigations using quantitative low and high contrast objects module serving as a references for each clinical image indicated that clinical images should be compressed according to their diagnostic content. Certain images can receive greater rates of compression than others and still retain all diagnostic information. The use of image compression (lossy compression) make Picture Archiving and Communication System (PACS) a more economically viable alternativeto analog film-based system by reducing the bite sized required to store and represent images while maintaining relevant diagnostic information. In this thesis, the results of the statistical tests showed thet there was no significant difference between the original and reconstructed images for compression rations up to 20 : 1 for Ct images (brain, chest and abdomen) and 50 : 1 for digitized X-ray images (Chest and abdomen), at a 95% of confidence level. Moreover, the results of ROC analysis suggested that compression ratios using LuraWave Smart Compression software for digitized X-ray images might be as high as = 100 : 1, without adverse effect on clinical diagnostic performance.